Search Images Maps Play YouTube News Gmail Drive More »
Sign in
Screen reader users: click this link for accessible mode. Accessible mode has the same essential features but works better with your reader.

Patents

  1. Advanced Patent Search
Publication numberUS8001104 B2
Publication typeGrant
Application numberUS 12/194,357
Publication dateAug 16, 2011
Filing dateAug 19, 2008
Priority dateNov 17, 2004
Also published asCN1776675A, CN1776675B, US7444360, US20060106898, US20080313416
Publication number12194357, 194357, US 8001104 B2, US 8001104B2, US-B2-8001104, US8001104 B2, US8001104B2
InventorsRhea Reyes Frondozo, Kurt Allen Lovrien, Richard Anthony Ripberger, Jacob Lee Sheppard, Michael P. Vageline
Original AssigneeInternational Business Machines Corporation
Export CitationBiBTeX, EndNote, RefMan
External Links: USPTO, USPTO Assignment, Espacenet
Method, system and program for storing and using metadata in multiple storage locations
US 8001104 B2
Abstract
Provided are a method, system, and program for storing and using metadata in multiple storage location. Signature data is stored in a system storage indicating a plurality of metadata copy locations, each locating identifying a storage device and a copy location within the storage device. Each location contains one copy of the metadata, wherein the metadata includes system configuration information. A copy of the signature data is stored with each copy of the metadata.
Images(9)
Previous page
Next page
Claims(21)
1. A method, comprising:
generating signature data indicating a plurality of metadata copy locations, each metadata copy location identifying one of a plurality of storage devices and a copy location within the storage device, wherein each copy location contains one copy of the metadata, wherein the metadata includes system configuration information;
storing a copy of the signature data with copies of the metadata in the storage devices indicated in the metadata copy locations of the signature data;
using the signature data to determine a copy location of valid metadata;
reading one copy of the valid metadata from the determined copy location in one of the storage devices to obtain system configuration information from the valid metadata to use to operate;
initiating an operation to alter the copy locations of the metadata copies in the storage devices storing the metadata;
determining new copy locations in the storage devices for the metadata;
updating each copy location of the signature data to indicate the new copy locations for the metadata; and
copying metadata and the updated signature data to any of the new copy locations that do not already include metadata and signature data.
2. The method of claim 1, wherein the location of the storage devices to store the copies of the metadata are selected to minimize a number of points of failure.
3. The method of claim 1, wherein the metadata copy locations indicated in the signature data comprise an ordered set, wherein the signature data is updated at the copy locations in an order according to the ordered set.
4. The method of claim 1, wherein the determination of the new copy locations is made in response to a change in the set of storage devices and wherein the determination of the new copy locations is performed in order to minimize a number of points of failure for the changed set of the storage devices.
5. The method of claim 1, wherein two processing systems are capable of accessing the metadata and the signature data in the storage devices, and wherein each processing system has a computer readable medium including a copy of the signature data.
6. The method of claim 5, further comprising:
receiving an update to the metadata; and
serially copying the update to the metadata in the storage devices according to an ordering of the copy locations indicated in the signature data.
7. The method of claim 5, further comprising:
receiving a request to read the metadata;
accessing the metadata in the copy locations until one valid copy of the metadata is found; and
updating copies of the metadata determined to be invalid with the accessed valid copy of the metadata.
8. A system, comprising:
a plurality of storage devices;
a processing system having a system storage and in communication with the storage devices; and
a computer readable medium including code executed by the processing system to perform operations comprising:
generating signature data indicating a plurality of metadata copy locations, each metadata copy location identifying one of a plurality of storage devices and a copy location within the storage device, wherein each copy location contains one copy of the metadata, wherein the metadata includes system configuration information;
storing a copy of the signature data with copies of the metadata in the storage devices indicated in the metadata copy locations of the signature data;
using the signature data to determine a copy location of valid metadata;
reading one copy of the valid metadata from the determined copy location in one of the storage devices to obtain system configuration information from the valid metadata to use to operate;
initiating an operation to alter the copy locations of the metadata copies in the storage devices storing the metadata;
determining new copy locations in the storage devices for the metadata;
updating each copy location of the signature data to indicate the new copy locations for the metadata; and
copying metadata and the updated signature data to any of the new copy locations that do not already include metadata and signature data.
9. The system of claim 8, wherein the locations of the storage devices to store the copies of the metadata are selected to minimize a number of points of failure.
10. The system of claim 8, wherein the copy locations within the signature data comprises an ordered set, wherein the signature data is updated at the copy locations in an order according to the ordered set.
11. The system of claim 10, wherein the determination of the new copy location is made in response to a change in the set of storage devices and wherein the determination of the new location is performed in order to minimize a number of points of failure for the changed set of the storage devices.
12. The system of claim 8, wherein the processing system comprises a first processing system, further comprising:
a second processing system, wherein both processing systems are capable of accessing the metadata and the signature data in the storage devices, and wherein each processing system has a local storage including a copy of the signature data.
13. The system of claim 12, wherein the code further causes operations comprising:
receiving an update to the metadata; and
serially copying the update to the metadata in the storage devices according to an ordering of the copy locations indicated in the signature data.
14. The system of claim 12, wherein the code further causes operations comprising:
receiving a request to read the metadata;
accessing the metadata in the copy locations until one valid copy of the metadata is found; and
updating copies of the metadata determined to be invalid with the accessed valid copy of the metadata.
15. An article of manufacture comprising at least one of a hardware device having hardware logic and a computer readable storage medium having executable code to communicate with a plurality of storage devices and a system storage and to cause operations to be performed, the operations comprising:
generating signature data indicating a plurality of metadata copy locations, each metadata copy location identifying one of a plurality of storage devices and a copy location within the storage device, wherein each copy location contains one copy of the metadata, wherein the metadata includes system configuration information;
storing a copy of the signature data with copies of the metadata in the storage devices indicated in the metadata copy locations of the signature data;
using the signature data to determine a copy location of valid metadata;
reading one copy of the valid metadata from the determined copy location in one of the storage devices to obtain system configuration information from the valid metadata to use to operate;
initiating an operation to alter the copy locations of the metadata copies in the storage devices storing the metadata;
determining new copy locations in the storage devices for the metadata;
updating each copy location of the signature data to indicate the new copy locations for the metadata; and
copying metadata and the updated signature data to any of the new copy locations that do not already include metadata and signature data.
16. The article of manufacture of claim 15, wherein the location of the storage devices to store the copies of the metadata are selected to minimize a number of points of failure.
17. The article of manufacture of claim 15, wherein the copy locations within the signature data comprises an ordered set, wherein the signature data is updated at the copy locations in an order according to the ordered set.
18. The article of manufacture of claim 15, wherein the determination of the new copy location is made in response to a change in the set of storage devices and wherein the determination of the new copy location is performed in order to minimize a number of points of failure for the changed set of the storage devices.
19. The article of manufacture of claim 15, wherein two processing systems are capable of accessing the metadata and the signature data in the storage devices, and wherein each processing system has a computer readable medium including a copy of the signature data.
20. The article of manufacture of claim 19, wherein the operations further comprise:
receiving an update to the metadata; and
serially copying the update to the metadata in the storage devices according to an ordering of the copy locations indicated in the signature data.
21. The article of manufacture of claim 19, wherein the operations further comprise:
receiving a request to read the metadata;
accessing the metadata in the copy locations until one valid copy of the metadata is found; and
updating copies of the metadata determined to be invalid with the accessed valid copy of the metadata.
Description
RELATED APPLICATIONS

This patent application is a continuation of U.S. patent application Ser. No. 10/991,664, filed on Nov. 17, 2004, which patent application is incorporated herein by reference in its entirety.

BACKGROUND OF THE INVENTION

1. Field of the Invention

The present invention relates to a method, system, and program for storing and using metadata in multiple storage locations.

2. Description of the Related Art

In certain computing environments, multiple host systems may communicate with multiple control units, such as an IBM Enterprise Storage Server (ESS)®, for data in a storage device managed by the ESS receiving the request, providing access to storage devices, such as interconnected hard disk drives through one or more logical paths. (IBM and ESS are registered trademarks of IBM). The interconnected drives may be configured as a Direct Access Storage Device (DASD), Redundant Array of Independent Disks (RAID), Just a Bunch of Disks (JBOD), etc. The control units maintain critical metadata in local storage that is needed to determine the configuration of the system. If the critical metadata is corrupted, invalid or unavailable, then the control unit will fail and may have to be completely reconfigured. The critical metadata is typically hidden from the host system and only available to the control unit operating system.

SUMMARY

Provided are a method, system, and program for storing and using metadata in multiple storage location. Signature data is stored in a system storage indicating a plurality of metadata copy locations, each locating identifying a storage device and a copy location within the storage device. Each location contains one copy of the metadata, wherein the metadata includes system configuration information. A copy of the signature data is stored with each copy of the metadata.

BRIEF DESCRIPTION OF THE DRAWINGS

FIG. 1 illustrates an embodiment of a computing environment in which embodiments are implemented.

FIG. 2 illustrates an embodiment of signature data information.

FIG. 3 illustrates an embodiment of information in a metadata record.

FIGS. 4-9 illustrate embodiments of operations to manage and use metadata and signature data.

DETAILED DESCRIPTION

FIG. 1 illustrates an embodiment of a computing environment in which aspects of the invention are implemented. One or more hosts 2 communicate Input/Output (I/O) requests directed to storage devices 4 a, 4 b . . . 4 n to a control unit 6, where the control unit 6 manages I/O access to the storage devices 4 a, 4 b . . . 4 n. In one embodiment, the control unit 6 is comprised of two systems 8 a, 8 b, each including a processor 10 a, 10 b, a cache 12 a, 12 b, and a local storage 14 a, 14 b. Each system 6 a, 6 b may be on separate power boundary. The two systems 8 a, 8 b may cooperate as a redundancy pair to perform the operation of the control unit such that in the event of a failure of either system, the remaining system takes over the operation of the control unit in a non-disruptive manner. The systems 6 a, 6 b may be assigned to handle I/O requests directed to specific volumes configured in the storage devices 4 a, 4 b . . . 4 n. The systems 6 a, 6 b communicate with the storage devices 4 a, 4 b . . . 4 n over a device network 16, which may comprise a local area network (LAN), storage area network (SAN), bus interface, serial interface, etc. The processors 8 a, 8 b execute I/O code 18 a, 18 to perform I/O and metadata management operations described herein. The local storage 14 a, 14 b may comprise a memory or a memory that is loaded from a non-volatile storage device.

The control unit 6 may comprise any type of server, such as an enterprise storage server, storage controller, etc., or other device used to manage I/O requests to attached storage devices 4 a, 4 b . . . 4 n, where the storage devices may comprise storage devices known in the art, such as interconnected hard disk drives (e.g., configured as a DASD, RAID, JBOD, etc.), magnetic tape, optical disks, electronic memory, etc. The hosts 2 may communicate with the control units 6 over a network (not shown), such as a Local Area Network (LAN), Storage Area Network (SAN), Wide Area Network (WAN), wireless network, etc. Alternatively, the hosts 2 may communicate with the control unit 6 over a bus interface, such as a Peripheral Component Interconnect (PCI) bus or serial interface. The processors 10 a, 10 b may communicate with each other over a connection 20 to handle failover or fallback.

The systems 8 a, 8 b maintain signature data 22 a, 22 b in their local storage 14 a, 14 b that indicates the location of copies of metadata 24 a, 24 b, 24 c in the storage devices 4 a, 4 b, 4 c. Copies of the signature data 26 a, 26 b, 26 c are also maintained with the metadata 24 a, 24 b, 24 c. The copies of the metadata 24 a, 24 b, 24 c may comprise critical metadata including critical configuration information needed for the control unit 6 to operate. If the critical metadata becomes corrupted or is unavailable, then the location and configuration of volumes on the storage devices may be lost and the control unit 6 may need to be reconfigured. The storage devices 4 a, 4 b, 4 c further include user data 28 a, 28 b, 28 c the hosts 2 access.

There may be more than the three storage devices 4 a, 4 b, 4 c shown and certain storage devices may not include copies of metadata and signature data. Further, one storage device may include multiple copies of the signature data and metadata. In one embodiment, three copies of metadata and signature data are stored in one or more storage devices. In alternative embodiments, a different number of copies of metadata and signature data may be maintained.

FIG. 2 illustrates an embodiment of information included with the signature data 22 a, 22 b, 26 a, 26 b, 26 c, system identifiers (IDs) 50 indicating the systems, e.g., 8 a, 8 b, for which the metadata identified by the signature data is intended; metadata locations 52 which identify one or more locations in the storage devices 4 a, 4 b, 4 c where a copy of metadata is to be found, each metadata location identifying a storage device 4 a, 4 b, 4 c and, if more than one copy of metadata can be located on a storage device 4 a, 4 b, 4 c, the identifier of the copy location on the identified storage device 4 a, 4 b, 4 c; and a generation number 54 incremented whenever the contents of the signature is modified due to a change in the metadata locations. The metadata 24 a, 24 b, 24 c may be stored at a predesignated location in the storage devices, known to the I/O code 18 a, 18 b, such that identifying the storage devices in the storage device IDs 54 provides sufficient information to locate the metadata 24 a, 24 b, 24 c and signature data 26 a, 26 b, 26 c in the storage devices 4 a, 4 b . . . 4 n.

Alternatively, there may be pointer metadata on the storage device at a location known to the I/O code that further identifies the locations of the copies of metadata on the storage device such that identifying the storage devices in the storage device IDs 54 and the identifier of a copy number on each storage device provides sufficient information to locate the pointer metadata on that storage device and the pointer metadata provides the location of the metadata copies 24 a, 24 b, 24 c and their signature data 26 a, 26 b, 25 26 c on the storage devices 4 a, 4 b . . . 4 n. If the storage device is a RAID array, each incarnation of a RAID array has a unique identifier. In one embodiment, the set of identified metadata locations in the signature data is ordered such that there is a 1st through Nth copy indicated. The ordering may be used to control algorithms used to validate signatures and access metadata as described subsequently.

FIG. 3 illustrates an embodiment of information included with a metadata record 70. In one embodiment, the metadata 24 a, 24 b, 24 c may be comprised of multiple metadata records. Each metadata record 70 may include a first and second generation numbers 72, 74 indicating a number of times the metadata record is updated, the actual metadata 76, and a error correction code (ECC) or a longitudinal redundancy check (LRC) code used to ensure the metadata record has not become corrupted.

The operations to use and manage metadata described in FIGS. 4, 5, 6, and 8 may be performed in a control unit that includes only one system or processor complex, i.e., no redundant hardware, or a control unit including two systems 8 a, 8 b such as shown in FIG. 1.

In one embodiment, there may be N candidate storage devices to store a copy of metadata on, each with M possible locations to store a copy of the metadata. As such there are N×M potential locations for the metadata to be stored. The control unit attempts to maintain K valid copies of the metadata, If K>N×M, then only N×M copies can be maintained. In one embodiment, K>=2 and N×M>=2 such that there are at least two copies of the metadata that are available to the control unit at any given time. Having at least two copies makes it possible to recover from a media error that corrupts a portion of one of the copies of the metadata. In another embodiment, K=3 and M=2 such that even when there is only one storage device on which to store metadata, there are at least two copies available on that storage device and when there are at least two storage devices on which to store metadata, there are three copies available, with at least two copies on independent storage devices.

FIG. 4 illustrates an embodiment of operations implemented in the I/O code 18 a and/or 18 b (the operations of FIG. 4 may be implemented on a control unit having one processor/system or multiple processors/system) executed by the processors 10 a, 10 b to provide an initial copy of the metadata. Upon initiating operations (at block 100) to select the optimal homes for metadata copies, the I/O code 18 a, 18 b makes (at block 102) an initial selection of one or more storage devices 4 a, 4 b, 4 c to store metadata to minimize the number of points of failure based on the current available storage devices by preferring to place 20 copies of the metadata in storage devices in different failure boundaries (e.g. on independent storage devices, on storage devices with independent access paths, on storage devices with independent power boundaries, etc.) or in locations where greater reliability is provided, (e.g. on higher reliability storage devices, on arrays with better RAID redundancy, on RAID arrays that are not degraded due to failing devices, on storage devices with more independent access paths, etc.).

As the set of available storage devices changes, the set of optimum metadata homes is reevaluated and the placement of the copies of the metadata is rearranged as required to keep the copies on a set of optimum homes. The evaluation of the set of optimum homes must be performed whenever the available set of metadata homes changes. This set may change as a result of adding or deconfiguring storage devices or as result of a storage device failure or repair. An initialized copy of metadata is created when the first storage device becomes available. As additional homes are added or removed, one or more existing copies of metadata may be moved from an existing home to a new home. The value of the signature data that is applicable changes each time the set of metadata homes used to store the metadata copies is modified. The generation number 54 (FIG. 2) of the signature data is incremented each time the signature data changes.

When a new copy of metadata is to be created or an existing copy of metadata is to be removed (or has become inaccessible), the signature data for the control unit must be updated. In the case of migrating a metadata copy from an existing home (storage device 4 a, 4 b, 4 c) to a more optimal home, the existing copy can be removed and then a new copy created.

FIG. 5 illustrates an embodiment of operations to remove a copy and create a new copy in order to update the signature data 50 when the configuration of the available storage devices 4 a, 4 b, 4 c changes, thereby causing a readjustment to select the optimal homes (storage devices 4 a, 4 b, 4 c . . . 4 n) for the metadata to minimize the number of points of failure based on the current available storage devices as discussed above. Upon initiating (at block 150) an operation to alter the homes (storage devices 4 a, 4 b, 4 c) for the metadata, from the new storage device configuration, the I/O code 18 a, 18 b determines (at block 152) an optimal number and location of storage devices 4 a, 4 b, 4 c for the metadata 70 to minimize the number of single points of failure. The I/O code performs operations 156 through 164 to remove the existing copy of the metadata 70. A determination is made (at block 154) of the new signature data value 22 a, 22 b, 26 a, 26 b, 26 c for the new configuration of metadata homes (optimal storage device 4 a, 4 b, 4 c selection). I/Os are quiesced (at block 156) to the set of metadata copies 24 a, 24 b, 24 c. The signature data value 22 a, 22 b, 26 a, 26 b, 26 c are set (at block 158) to zero on the copy being removed and the signature data values 26 a, 26 b, 26 c on the newly selected optimal storage devices 4 a, 4 b, 4 c are updated (at block 160) in the order indicated in the new signature data. The signature data values 22 a, 22 b on the local storage 14 a, 14 b of each system 8 a, 8 b are also updated (at block 162) and I/O resumes (at block 164) to the set of metadata copies 24 a, 24 b, 24 c.

The I/O code performs operations 166 through 176 to add a new copy of the metadata 70. The I/O code 18 a, 18 b determines (at block 166) the new signature data value for the new configuration of metadata homes and copies (at block 168) all metadata to the new homes, maintaining any updates to the metadata across all copies as they occur. The I/O code 18 a, 18 b quiesces (at block 170) I/O to the set of metadata copies 24 a, 24 b, 24 c and updates (at block 172) the signature data 26 a, 26 b, 26 c on the storage devices 4 a, 4 b, 4 c for each metadata home in the order found in the new signature data. The signature data 22 a, 22 b on the local storage 14 a, 14 b of each system 8 a, 8 b is also updated (at block 174) and I/O then resumes (at block 176) to the set of metadata copies 24 a, 24 b, 24 c.

In the embodiment of FIG. 5, the update of the signatures is done while I/O to the metadata copies is suppressed such that the transition to the new signatures happens as an atomic operation with all copies having the same metadata at the point the signatures are modified. At the end of the signature update process, all copies of the signature should be current on all local and storage device copies and any storages without a metadata copy should not have a signature provided that all these signature locations are accessible. However, in cases where there is only one processor online or where a storage device is inaccessible, older generations of signature may be left in place. If the update process is disrupted by a power loss, the signatures may be left in a transitional state.

The signatures are updated in a specific order so that other algorithms can be defined to resolve a valid data signature under reset conditions as described subsequently.

When one or both systems 8 a, 8 b of the control unit 6 power up and begin the operation of the control unit 6, the system(s) 8 a, 8 b must first locate the valid copies of the current metadata 24 a, 24 b, 24 c. A copy of the signature data 26 a, 26 b, 26 c is maintained with each copy of the metadata and is also stored locally, i.e., in local storage 14 a, 14 b, on each system 8 a, 8 b. The process of locating the current metadata is complicated by the fact that: not all the storage devices may be accessible to the operational system(s) due to failures; the signatures may have been in the middle of being updated when a power loss occurred; one of the systems may not have been operational and its local signature data may be out of date with any changes to the location of the metadata made by the other operational server; either one or both of the systems 8 a, 8 b may be operational at the point of the power up.

The I/O code 18 a, 18 b may validate that the current copies of metadata 24 a, 24 b, 24 c have been located with a certain degree of confidence. The signature data 22 a, 22 b, 26 a, 26 b, 26 c determines the location of the currently valid metadata 24 a, 24 b, 24 c, so that the I/O code 18 a, 18 b need only determine the validity of the signature data 26 a, 26 b, 26 c. If valid signature data is not determined, then the control unit 6 does not go online to the attached hosts 2 and calls for service to let a service person attempt to sort out what copy if any of the metadata should be used. If at least one valid copy of metadata 24 a, 24 b, 24 c can be located, then the system(s) 8 a, 8 b make additional copies of the metadata as previously described so that that the appropriate number of valid copies exist. The signature data 26 a, 26 b, 26 c in the storage devices 4 a, 4 b, 4 c in the local signature data 22 a, 22 b is updated to reflect the current valid copies once the location of one or more copies is determined. The algorithm to determine a valid signature may depend on the procedure used to update the copies of the signature data when the applicable signature changes for the control unit.

FIGS. 6 and 7 illustrate one embodiment of operations to validate a copy of signature data value 22 a, 22 b, 26 a, 26 b, 26 c when powering-on or as part of an initial microcode load (IML) operation. At block 200, one system, e.g., 8 a, performs an initialization, such as a power-on or IML, after a resetting condition. If (at block 202) the other system, e.g., 8 b, is operational and if (at block 204) there is signature data 22 a, 22 b in both systems 8 a, 8 b, then the two signature data values 22 a, 22 b are compared (at block 206). If (at block 208) they both match and if (at block 210) at least one storage device signature data 26 a, 26 b, 26 c both: (1) matches the local storage signature data 22 a, 22 b or have a generation number 54 one greater than the local storage signature data 22 a, 22 b and (2) has at least one storage device in common with the local storage signature data 22 a, 22 b, then control proceeds to block 224 in FIG. 7. Otherwise, from the no branch of block 208 or 210, the I/O code 18 a determines (at block 212) whether there is one storage device signature data 26 a, 26 b, 26 c having a generation number 54 equal to or one greater than the local storage signature data 22 a, 22 b having the latest generation number and both of these local and storage device signature data identify the storage device 4 a, 4 b, 4 c having the generation number equal or greater, then control proceeds to block 224 in FIG. 7. Otherwise, from the no branch of block 212, a the control unit 6 does not go online (at block 214 and service may be called.

If (at block 202) the other system, e.g., 8 b, is not operational or if (at block 204) only one system 8 a or 8 b has signature data 22 a or 22 b, then control proceeds to block 220 in FIG. 7. At block 220, the I/O code 18 a scans all storage devices 4 a, 4 b, 4 c for signature data 26 a, 26 b, 26 c for metadata homes, keeping track of each signature and its location. If (at block 222) there are two matching signature data values 26 a, 26 b, 26 c on storage devices 4 a, 4 b, 4 c with a generation number 54 greater than any other signature data value (local storage 14 a or storage device 4 a, 4 b, 4 c), then the storage device signature data 26 a, 26 b, 26 c with the greater generation number is validated (at block 224). The I/O code 18 a updates (at block 226) any signature data values 22 a, 22 b s in local storage 14 a, 14 b and on the storage devices 4 a, 4 b, 4 c indicated (in field 54 of FIG. 2) in the validated signature data that are not the same as the validated signature data. The signature data 26 a, 26 b, 26 c on any storage devices 4 a, 4 b, 4 c that are not identified by the validated signature data values are zeroed (at block 228). If the control unit 6 is currently offline (i.e. a power on for the first operational processor), then the I/O code 18 a synchronizes (at block 230) copies of metadata by copying from the first copy through to the Jth copy to remove any partial updates from prior power off. The I/O code 18 a further starts (at block 232) the processes to replicate or migrate metadata copies to optimum homes as required and then goes online (at block 234).

From the no branch of block 222, if (at block 236) there is one storage device having signature data 26 a, 26 b, 26 c one generation number 54 less than another storage device signature data and the signature data for both storage devices identifies the storage devices, then control proceeds to block 224 to validate the signature data value 26 a, 26 b, 26 c with the latest generation number. From the no branch of block 236, if (at block 238) there is one storage device having signature data 26 a, 26 b, 26 c one generation number 54 greater than or equal to local storage 14 a signature data value 22 a, 22 b and the signature data value for the storage device having the greatest generation number and for the storage system identifies the storage device 4 a, 4 b, 4 c having the signature data value 26 a, 26 b, 26 c with the latest generation number 54, then control proceeds to block 224 to validate the storage system signature data 26 a, 26 b, 26 b having the latest generation number one generation greater than the local signature data 22 a, 22 b. Otherwise, from the no branch of block 238, control proceeds to block 240 to not go online and call for service. Other embodiments of the algorithm may choose different results (online or offline) for the conditions checked or might define additional conditions to check depending on the level of confidence desired in the determination of the validity of the signature.

The signature data value identifies an ordered set of metadata homes with valid copies of the metadata. When any portion (i.e. a record) of the metadata is updated, the updates are made in this order to each metadata copy and the update is not considered complete until all copies have been updated as described subsequently. In the case of a power loss or certain failures that cause discontinuance of the operation of all systems in the control unit, it is possible that not all copies of metadata get updated. To prevent the metadata copies from being out of sync, at power on time, the systems copy the data from the first copy of the metadata to the remaining copies (at block 230 in FIG. 7). Any errors detected on the first copy of the metadata are recovered by using the valid data available on the next successive copy of the metadata. If an update was not in progress on the data in error, then all copies should have had the same data and the recovery restores the current data to all copies. If a partial update was in progress, then taking the data from the next valid successive copy either restores all copies to the version of the metadata that was intended to be updated or restores all copies to the version of the metadata that existed prior to the update. In the first case, the partial update is completed. In the later case, the partial update is nullified. In either case the metadata is restored to a consistent state since the partial update was disrupted before it was complete and the system design must be able to handle that an incomplete partial update either occurred or did not occur.

When any portion (i.e. a record) of the metadata is updated, the updates are made in this order to each metadata copy and the update is not considered complete until all copies have been updated. The updates are serialized to the multiple copies of the metadata so that a power loss does not end up partially updating all three copies resulting in the loss of the metadata in all copies. Read and write accesses to the same portion (i.e. record) of the metadata are generally handled such that there are no read operations allowed while a write operation is in progress. In this case, a read may be issued to any available copy of the metadata (i.e. all copies contain the same data assuming they were synchronized at power on). A read failure (e.g. media error) on any copy is recovered by reading data from another valid copy of metadata until a valid copy of the portion of metadata is found. Once found any copies in error are re-written with valid data. If no valid copy is found and error is returned to the accessor of the data indicating that the data has been lost. In cases where concurrent read and write accesses might occur to the same portion of metadata, the reads are issued in the order that updates are performed so that the latest copy of the metadata is read. If a failure is detected on either a read or a write access that is indicative of a failure of the entire copy of metadata, then that copy is considered inaccessible, removed from the signature data, and a new copy is created in another metadata location, if any are available.

FIG. 8 illustrates an embodiment of operations to update a single metadata record 70 (FIG. 3) in one of the copies 24 a, 24 b, 24 c of the metadata implemented in the I/O code 18 a, 18 b. (The operations of FIG. 8 may be implemented on a control unit having one processor/system or multiple systems 8 a, 8 b). Upon initiating operations (at block 300) to update a specified metadata record 70, the update to the specified metadata record is written (at block 302) to each copy of the metadata 24 a, 24 b, 24 c in the at least one storage device 4 a, 4 b, 4 c by performing the operations at blocks 304, 306, and 308 on one record copy at a time. At block 304, the I/O code 18 a increments the first generation number 72 and writes the first block to the metadata record 70. All blocks except the first and last block are written (at block 306) to the metadata record 70. The second generation number 76 is incremented (at block 308), the LRC/CRC calculated for the entire metadata record, and the last block is written to the metadata record 70.

FIG. 9 illustrates an embodiment of operations to read a single metadata record 70 (FIG. 3) in one of the copies 24 a, 24 b, 24 c of the metadata implemented in the I/O code 18 a, 18 b. (The operations of FIG. 9 may be implemented on a control unit having one processor/system or multiple systems 8 a, 8 b). At block 350, an operation is initiated to read a specified metadata record 70 on a specified copy of the metadata 24 a, 24 b, 24 c comprising a plurality of metadata records, where there are a first 70 and second 76 (FIG. 3) generation numbers associated with each metadata record 70 indicating a number of times the metadata record has been updated and an ECC or LRC check character. The I/O code 18 a attempts (at block 352) to read the specified metadata record 70 on one storage device 4 a, 4 b, 4 c providing a home for the specified copy of the metadata 70. If (at block 354) there is an access error, then the access error is returned (at block 356) to a higher level accessor function. If (at block 354) there is no error, then the first 72 and second 76 generation numbers are checked (at block 358). If (at block 360) they do not match, then a read error is returned (at block 362) to a higher level. If (at block 364) an LRC or uncorrectable CRC error is detected on the metadata record, then a read error is returned to a higher level. If no errors were detected, the metadata record has been validated and is returned to the accessor function.

Additional Embodiment Details

The described embodiments may be implemented as a method, apparatus or article of manufacture using standard programming and/or engineering techniques to produce software, firmware, hardware, or any combination thereof. The term “article of manufacture” as used herein refers to code or logic implemented in hardware logic (e.g., an integrated circuit chip, Programmable Gate Array (PGA), Application Specific Integrated Circuit (ASIC), etc.) or a computer readable medium, such as magnetic storage medium (e.g., hard disk drives, floppy disks, tape, etc.), optical storage (CD-ROMs, optical disks, etc.), volatile and non-volatile memory devices (e.g., EEPROMs, ROMs, PROMs, RAMs, DRAMs, SRAMs, firmware, programmable logic, etc.). Code in the computer readable medium is accessed and executed by a processor. The code in which preferred embodiments are implemented may further be accessible through a transmission media or from a file server over a network. In such cases, the article of manufacture in which the code is implemented may comprise a transmission media, such as a network transmission line, wireless transmission media, signals propagating through space, radio waves, infrared signals, etc. Thus, the “article of manufacture” may comprise the medium in which the code is embodied. Additionally, the “article of manufacture” may comprise a combination of hardware and software components in which the code is embodied, processed, and executed. Of course, those skilled in the art will recognize that many modifications may be made to this configuration without departing from the scope of the present invention, and that the article of manufacture may comprise any information bearing medium known in the art.

In one embodiment, the first generation number may be in a first sector of each metadata record and the second generation number in a last sector of the metadata record.

In embodiments having redundant hardware, e.g., systems 8 a, 8 b, one system 8 a or 8 b may be designated the master to control the storage devices 4 a, 4 b, 4 c in which copies of the metadata 24 a, 24 b, 24 c are stored. In one embodiment, the master system 8 a or 8 b may assign copies of the metadata 24 a, 24 b, 24 c to the systems 8 a, 8 b, such that the systems 8 a, 8 b manage the locking of their assigned metadata copies 24 a, 24 b, 24 c.

FIGS. 2 and 3 show certain information included in the signature data and metadata records. In alternative embodiments, this information may be stored in different data structures having different formats and information than shown.

The illustrated operations of FIGS. 4-8 show certain events occurring in a certain order. In alternative embodiments, certain operations may be performed in a different order, modified or removed. Moreover, steps may be added to the above described logic and still conform to the described embodiments. Further, operations described herein may occur sequentially or certain operations may be processed in parallel. Yet further, operations may be performed by a single processing unit or by distributed processing units.

The foregoing description of various embodiments of the invention has been presented for the purposes of illustration and description. It is not intended to be exhaustive or to limit the invention to the precise form disclosed. Many modifications and variations are possible in light of the above teaching. It is intended that the scope of the invention be limited not by this detailed description, but rather by the claims appended hereto. The above specification, examples and data provide a complete description of the manufacture and use of the composition of the invention. Since many embodiments of the invention can be made without departing from the spirit and scope of the invention, the invention resides in the claims hereinafter appended.

Patent Citations
Cited PatentFiling datePublication dateApplicantTitle
US5325497Mar 29, 1990Jun 28, 1994Micro Technology, Inc.Method and apparatus for assigning signatures to identify members of a set of mass of storage devices
US5909691Dec 20, 1996Jun 1, 1999Compaq Computer CorporationMethod for developing physical disk drive specific commands from logical disk access commands for use in a disk array
US5960194Jul 17, 1996Sep 28, 1999International Business Machines CorporationMethod for generating a multi-tiered index for partitioned data
US6041334 *Apr 13, 1999Mar 21, 2000International Business Machines CorporationStorage management system with file aggregation supporting multiple aggregated file counterparts
US6078999Dec 29, 1998Jun 20, 2000Microsoft CorporationRecovering from a failure using a transaction table in connection with shadow copy transaction processing
US6233589 *Jul 31, 1998May 15, 2001Novell, Inc.Method and system for reflecting differences between two files
US6266740Jul 31, 1998Jul 24, 2001Emc CorporationDevice sequencing and placement protection using a digital signature
US6499039 *Sep 23, 1999Dec 24, 2002Emc CorporationReorganization of striped data during file system expansion in a data storage system
US6505209 *Nov 2, 1999Jan 7, 2003Monkeymedia, Inc.Poly vectoral reverse navigation
US6745147Oct 4, 2001Jun 1, 2004International Business Machines CorporationData processing system, method, and product for automatically tracking insertions of integrated circuit devices
US6785688 *Jun 8, 2001Aug 31, 2004America Online, Inc.Internet streaming media workflow architecture
US6889309 *Apr 15, 2002May 3, 2005Emc CorporationMethod and apparatus for implementing an enterprise virtual storage system
US7130956 *Feb 10, 2004Oct 31, 2006Sun Microsystems, Inc.Storage system including hierarchical cache metadata
US7444360Nov 17, 2004Oct 28, 2008International Business Machines CorporationMethod, system, and program for storing and using metadata in multiple storage locations
US20020024963Jul 5, 2001Feb 28, 2002Andrew ReeveBuffer management
US20020053008May 3, 2001May 2, 2002International Business Machines CorporationVerifying primary and backup copies of vital information for a processing system employing a pseudo-fixed reference identifier
US20020143976 *Mar 7, 2002Oct 3, 2002N2Broadband, Inc.Method and system for managing and updating metadata associated with digital assets
US20030182322 *Mar 19, 2002Sep 25, 2003Manley Stephen L.System and method for storage of snapshot metadata in a remote file
US20030195903Apr 23, 2003Oct 16, 2003Manley Stephen L.System and method for asynchronous mirroring of snapshots at a destination using a purgatory directory and inode mapping
US20030221165 *May 22, 2002Nov 27, 2003Microsoft CorporationSystem and method for metadata-driven user interface
US20040025008 *Aug 2, 2002Feb 5, 2004The Boeing CompanySystem, method and apparatus for securing network data
US20040039892Aug 23, 2002Feb 26, 2004Goldschmidt Marc A.System and method of accessing vital product data
US20040044863Aug 30, 2002Mar 4, 2004Alacritus, Inc.Method of importing data from a physical data storage device into a virtual tape library
US20040088333Jan 25, 2002May 6, 2004David SidmanApparatus method and system for tracking information access
US20040098547Jun 30, 2003May 20, 2004Yuval OfekApparatus and methods for transferring, backing up, and restoring data in a computer system
US20040143713Jan 22, 2003Jul 22, 2004Niles Ronald S.System and method for backing up data
US20040167934Jan 7, 2004Aug 26, 2004Permabit, Inc., A Massachusetts CorporationHistory preservation in a computer storage system
US20040167935Jan 7, 2004Aug 26, 2004Permabit, Inc.History preservation in a computer storage system
US20040172501 *Aug 15, 2003Sep 2, 2004Hitachi, Ltd.Metadata allocation method in a storage system
US20040215644 *May 14, 2004Oct 28, 2004Edwards Robert ClairApparatus, method, and system for aggregated no query restore
US20040220791 *Jun 2, 2004Nov 4, 2004Interactual Technologies, Inc. A California CorporPersonalization services for entities from multiple sources
US20060095470 *Nov 4, 2004May 4, 2006Cochran Robert AManaging a file in a network environment
Non-Patent Citations
Reference
1Final Office Action IFW dated Sep. 7, 2007, pp. 1-14, for U.S. Appl. No. 10/991,664, by inventors R.R. Frondozo, et al.
2First Office Action IFW dated Apr. 17, 2007, pp. 1-23, for U.S. Appl. No. 10/991,664, by inventors R.R. Frondozo, et al.
3Notice of Allowance IFW dated Jun. 20, 2008, pp. 1-8, for U.S. Appl. No. 10/991,664, by inventors R.R. Frondozo, et al.
4Request for Continued Examination IFW dated Mar. 7, 2008, pp. 1-3, in response to Final Office Action IFW dated Sep. 7, 2007, pp. 1-14, for U.S. Appl. No. 10/991,664, by inventors R.R. Frondozo, et al.
5Response IFW dated Aug. 11, 2007, pp. 1-17, to Final Office Action IFW dated Sep. 7, 2007, pp. 1-14, for U.S. Appl. No. 10/991,664, by inventors R.R. Frondozo, et al.
6Response IFW dated Jun. 18, 2007, pp. 1-18, to First Office Action IFW dated Apr. 17, 2007, pp. 1-23, for U.S. Appl. No. 10/991,664, by inventors R.R. Frondozo, et al.
7Supplemental Response IFW dated Jan. 8, 2008, pp. 1-12, to Final Office Action IFW dated Sep. 7, 2007, pp. 1-14, for U.S. Appl. No. 10/991,664, by inventors R.R. Frondozo, et al.
Referenced by
Citing PatentFiling datePublication dateApplicantTitle
US8533478 *Oct 24, 2008Sep 10, 2013Hewlett-Packard Development Company, L. P.System for and method of writing and reading redundant data
US20100106974 *Oct 24, 2008Apr 29, 2010Aguilera Marcos KSystem For And Method Of Writing And Reading Redundant Data
Classifications
U.S. Classification707/706, 707/821
International ClassificationG06F17/30
Cooperative ClassificationG06F11/2094, G06F11/2056, G06F3/0614, G06F3/065, G06F3/067, Y10S707/99952, G06F11/1435, G06F17/30067, G06F11/2089
European ClassificationG06F17/30F, G06F11/14A8F